[View:http://imgur.com/a/soD8e:800:0]Jeppe.

Its playing now, you can come and pick it up
I have a slight dyslexia. I have trouble of following long threads and long emails.
So I have not read all the posts here.

The problem was driver transistor to the regulator, the one you replaced and most likely the problem is your own cooling invention

The cooling solution you have done is the cause.
The AC128 was shortcut, and did not provide proper level to the power transistor, which made power transistor to level out way to high  voltage. 

 ( based on how it looked like when I got it there is also a speculative reasons that , due to eventual capsel leakage it was short cut with your cooling solution , to the cables in the surrounding )

There is a reason why the orginal colling design is as it is.
The coppar plate rounding must round around the transitor perfect.

The way to to it is to desoilder the entire package of both transistors and the colling plate.

Dillen wrote some where, I cant remember, "lots of strong coffe and good coockies helps"

Now there were more problems. the PCB layout had loosen from one pin,  because of not proper earlier demounting, which of course broke when I repaired, so I had to make sad solutions.
So , the only thing I can teach you is that there are no short cuts, do it right ;) ;) ;)

Beoworld servers didnt allow me to upload images this night, so link is bellow
[View:http://imgur.com/a/soD8e:800:0]

What I did is I took out the entire package, tried my best to restore the rounding of the coppar, to my best and inserted a new AC128 and filled it with silicon paste to hopefully compensate for the not perfect rounding..

Sombody had wrinkled out the coppar rounding so it was crinkled, not easy to get perfect.

Keep the orginal, think orginal. 

Christian Christensen wrote:

Dillen wrote some where, I cant remember, "lots of strong coffe and good coockies helps"

The cookies has to be Danish

Thank you Christian!!!

I did try to be so careful when I made the heat sink. I even used thermal paste between the copper "cage" and the transistor. But I see, it didn't work. I was pretty sure myself that it would disperse heat enough, then perhaps it was shorted. I was thinking when I made it - "I hope the outer part of the transistor is properly insulated, otherwise I am screwed"..

So the AC128's actually get hot enough to need a proper heat sink. I wasn't sure of that either.

Even with Danish cookies and strong coffee, to remove that heat sink and fix it and put it back must have been lots of difficult work!

Looking forward very much to testing it and comparing it to the almost unrestored white Beomaster 1000 that I also have.

. Sören, As we can buy danish coockies in Sweden, beoing is easier in this country compared to other countries 😕

I am investigating the possibility to remake a few pieces of that front rail that holds the radio scale, between the keys, on the Beomaster 1000. It seems they often break. On one of mine it is broken, and I want to replace it. Now checking if they can be 3-D printed or molded. It seems it can be done.

Anyone here who is interested in having this spare part? In the photo is a typical broken rail.

Login to view image

Jeppe wrote:

Now the question is - how do you search for faulty germanium transistors?
I know how to measure silicon ones. But these germanium transistors give the oddest readings, they are all over the place and I don't know how to check if they are alright or not!

How do you guys do it on your germanium machines?😕

Germanium transistors are quite "bad" compared to silicon transistors - for example, they may have a large reverse leakage current that can mess up simple multimeter diode tests. Regular cheap multimeters with a hFE range may fail as well.

The best option would be a real transistor tester, if you do enough work to justify the cost - I haven't looked in a while, but I would expect that you could find a cheap but useable one from China nowadays. Or perhaps even build a small test bench:

link

(theory for this more extensively here:

Thanks Mika!
It might be a good idea to get me a transistor tester, for example the Peak DCA55 mentioned in the link you sent. It seems it tests germanium as well.

And they can be tested in-circuit?

I am determined to get this BM1000 in working order..

Jeppe wrote:

And they can be tested in-circuit?

Not really. And be careful when (de)soldering Ge transistors, they are very sensitive to heat. That's why the leads are often left quite long, and the old advice was to hold the wires with pliers while soldering so that most of the heat doesn't reach the actual component.

That Peak instrument looks nice, I might get myself one.

Oh dear, they must be desoldered to be checked?

Addition: I got an idea from youtube, it may be enough to desolder them to test them, you may not actually have to remove them from the board if you can desolder cleanly. This is an advantage as some of them sit in heatsinks and it means quite a lot of work to remove it all.